MARINE ECOLOGY PROGRESS SERIES Vol. 280: 227–238, 2004 Published October 14 Mar Ecol Prog Ser Feeding habits of pandalid shrimps in the Alboran Sea (SW Mediterranean): influence of biological and environmental factors Emanuela Fanelli*, Joan E. Cartes Institut de Ciéncies del Mar, Passeig Maritim de la Barceloneta 37–49, 08003 Barcelona, Spain ABSTRACT: The feeding habits of 4 species of Plesionika inhabiting the Alboran Sea (SW Mediter- ranean) at depths between 29 and 790 m were analyzed. Included within the trophic guild of non- migratory macroplankton feeders, Plesionika spp. preferably exploited mesopelagic resources (mainly euphausiids and Myctophidae). Deep-bathyal species (P. martia, P. gigliolii and P. edwardsi) fitted better within this guild, while the shallower living species P. heterocarpus, which inhabits the shelf-slope break, consumed benthic organisms (e.g. polychaetes, bivalves and tanaids). In a multi- species MDS analysis, diets of P. martia and P. heterocarpus were grouped as a function of depth. Lat- itude, phytoplankton pigment (PP) concentration recorded 4 wk before the sampling date (PP4) and depth were the explanatory variables of stomach fullness of Plesionika (all species: R2 = 0.674 in a multiple linear regression model, MLR). The abundance of P. martia also showed a positive signifi- cant correlation (Spearman r) with PP4. Possible causes of higher shrimp aggregations in P. martia could be related to reproductive processes because peaks of mature females were reported at similar periods in neighboring areas. In both dominant Plesionika species (P. martia and P. heterocarpus), MDS dimensions for diets were also correlated with PP4, although significant correlations with other PP data taken at different periods (from simultaneous PP data to PP6, 6 wk before sampling) were also found. These trends suggest that macroplankton-feeder Pandalidae may couple some aspects of their biology with blooms of plankton and ultimately with peaks of primary production. Prey size selection occurred among Plesionika spp., which suggested a hunting activity. Correlations between Plesionika spp. size (CL, mm) and prey size were positive, both for fish (R2 = 0.437) and for euphausi- ids (R2 = 0.440). Otolith length (mainly of Benthosema glaciale) and the size of euphausiid mandibles (Meganyctiphanes norvegica and Euphausia hemigibba) increased with predator size. Further evidence of active predation by deep pandalids, for example changes in the daily feeding activity of species, were evidenced for P. martia with higher foregut fullness at the beginning of the daytime period (from 06:00 to 08:00 h) related to predation on euphausiids. KEY WORDS: Pandalid shrimps · Diet · Feeding guilds · Primary production Resale or republication not permitted without written consent of the publisher INTRODUCTION culated for fish (Cartes & Maynou 1998). Pandalids are one of the most characteristic and diversified families Decapod crustaceans are a dominant taxon at slope of marine shrimps. The genus Plesionika (Decapoda, depths in tropical and subtropical regions, the case of Caridea) is widely distributed in tropical and subtropi- the deep Mediterranean (Cartes 1993c), with an im- cal areas (Zariquey Alvarez 1968, Crosnier & Forest portant contribution to the energy flow among the 1973). P. heterocarpus (Costa, 1871) is abundant in the highest trophic levels. Thus, in a preliminary energy mud assemblage of the shelf-slope break and the balance study performed in the Catalan Sea, food con- upper slope of the Alboran Sea (SW Mediterranean), sumption deduced for decapods was similar to that cal- while P. gigliolii (Senna, 1903), P. edwardsii (Brandt, *Email: [email protected] © Inter-Research 2004 · www.int-res.com 228 Mar Ecol Prog Ser 280: 227–238, 2004 1851) and P. martia (A. Milne Edwards, 1883) are dom- logical variables (e.g. sexual state, and density) were inant in bathyal communities of the upper and middle considered. slope. The Alboran Sea is characterized by the domi- nance of pandalid shrimps, which reach the highest values of abundance along the Iberian coasts. There is MATERIALS AND METHODS a tendency to shallower distribution of Plesionika spp. in the Alboran Sea compared with other study areas During cruise MEDITS-ES02 (11 to 17 May 2002), 27 (Carbonell & Abelló 1998). P. heterocarpus is dis- hauls were carried out between depths of 29 and 790 m tributed in the Alboran Sea from 47 to 467 m, with a along the coasts of the Alboran Sea (SW Medi- bimodal distribution and the highest abundance terranean) from the Strait of Gibraltar (ca. 36° 00’ N, between 100 and 150 m, and 200 and 300 m depth. P. 5° 50’ W) to the Cape of Gata (36° 38’ N, 2° 05’ W), over edwardsii occurs from 300 to 400 m, always with very a spatial scale of ca. 300 km. Plesionika shrimps were irregular and rather low densities. P. gigliolii is distrib- collected from trawl hauls between 69 and 715 m. uted from 248 to 600 m with peaks of abundance Shrimps were sampled with a bottom trawl (Bertrand between 300 to 500 m, whereas P. martia was found et al. 2000). The horizontal aperture of the gear was ca. almost exclusively on the middle slope of the Alboran 18 m and the vertical opening was ca. 2 m. The cod- Sea from 300 to 800 m (Carbonell & Abelló 1998). Sim- end mesh size was 10 mm. Haul duration was 0.5 h at ilar distributions and patterns of species substitution depths <200 m and 1 h at greater depths. with depth were found in neighboring areas (Cartes Abundance and biomass were calculated (N ind. km–2) et al. 1993, Maynou & Cartes 2000). after standardizing catches to swept area (km2). Dis- The trophic habits of some pandalid shrimps were tance and mouth trawl width were recorded using a reported by Lagardère (1972, 1977), Wienberg (1980), SCANMAR system attached to the trawl for each haul. Cartes (1993a,b, 1998) and Mura (1995). Plesionika A total of 240 specimens of Plesionika heterocarpus, species mainly exploit benthopelagic resources, prey- 71 of P. gigliolii, 21 of P. edwardsii and 231 of P. martia ing on euphausiids and mesopelagic decapods (e.g. were studied to establish the foregut fullness for Pasiphaea spp. and Sergestidae). They can be in- each species (see Table 1). Individuals were measured cluded in a trophic guild (Cartes 1998, Cartes et al. (carapace length, CL, in mm with an accuracy of 2002), with only some species exploiting other com- 0.1 mm) dorsally from the posterior edge of the eye partments (e.g. P. heterocarpus, which also consumes socket to the posterior edge of the cephalothorax using infauna, particularly polychaetes in the Bay of Biscay, a vernier caliper and weighed to the nearest 0.1 g. Lagardère 1977). Scavenging on fish remains were Information on gonadal development and ovigerous also reported, particularly in areas where the species is females was also collected. The wet weight of stomach not dominant (e.g. the case of P. martia) at the limit of contents was measured. Prey in stomach contents were its geographical distribution (Lagardère 1977). Size ascribed in the laboratory under a stereomicroscope segregation by coexisting Plesionika species having (10 to 40×) to the lowest possible taxonomic level. the same trophic guild has also been indicated in a pre- The following indexes were calculated (see Hyslop vious study performed in the Catalan Sea (Cartes 1980 for a review): (1) percentage of vacuity (empty 1994). stomachs / total stomachs × 100), (2) stomach fullness Although changes in temperature and phytoplank- (stomach weight / shrimp weight × 100), (3) frequency ton concentration have been related to the diversity of of occurrence (%F) of prey, (4) percentage of numeric fish communities (McClatchie et al. 1997, Jacob et al. abundance of prey (%N) and (5) percentage of wet 1998), such variables have been rarely considered at weight of prey (%W). The percentage of volume of the autoecological level (Cartes et al. 2004). Most feed- stomach contents was estimated using the subjective ing/dietary studies, therefore, lack an analysis of fac- points methods (Swynnerton & Worthington 1940) for tors influencing feeding activity (e.g. stomach fullness, the different prey-species or prey-groups. daily rations) and diet, which include prey/food avail- Trophic diversity was calculated on the diet of each ability as well as spatio-temporal short-scale changes species using the Shannon-Wiener index, H’ (Shannon in productivity and oceanographic conditions. Beyond & Weaver 1949), species richness, d (Margalef 1958) the description of the diets and feeding habits of the and the evenness index, J’ (Pielou 1966). 4 pandalid species occurring in the Alboran Sea, the General affinities between the samples (diets) were objective of this study was to analyze the possible established using multivariate cluster analysis (cf. influence of environmental and biological factors in Cartes & Abelló 1992) for Plesionika martia and the feeding and diet of Plesionika species. Physical P. heterocarpus, after a log-transformation of matrix (e.g. temperature and salinity), trophic (e.g. phyto- data, based on prey weight for each haul. Cluster plankton concentration and prey availability) and bio- analysis, using UPGMA (Pielou 1984) as the algorithm Fanelli & Cartes: Feeding habits of pandalid shrimps 229 of aggregation and 1 – r (the linear correlation of Pear- Relationships between the size of predators (Ple- son) as the distance, was used to classify the samples. sionika spp.) and their prey were also examined. Pan- On the basis of the similarity matrix generated, Multi- dalid prey are usually broken in the stomach; there- dimensional Scaling (MDS) techniques were applied fore, we considered maximum otolith length for fish for the ordination of samples (as a function of their prey (myctophids, mainly Benthosema glaciale) and mand- composition) along 2-dimensional axes. ible height for euphausiids (chiefly Euphausia hemi- We also calculated Spearman’s correlation coeffi- gibba and Meganyctiphanes norvegica).